Research On Grain Output Prediction Method Based On GA-BPNN

With the reform of the grain circulation system, grain circulation turns from the unified purchase and vendition to the open market. The country gradually take macro-control management to the grain market, a national grain reserve system is established. To strengthen the state grain reserve management of grain storage business, computer management of research and development is particularly important. The research and development of rain output prediction can strengthen the enterprises internal management of grain industry, and make scientific prediction of the current economic situation by quantitative analysis.This paper establishes a grain output prediction model based on Neural Network. The grain output data released by the State Statistics Bureau is used as training samples and test samples. The prediction by BP algorithm and the BP algorithm optimized by genetic algorithm in grain output prediction model are compared. The paper demonstrates BP algorithm optimized by GA has faster converging speed and higher accurate prediction than BP algorithm.Therefore, it can draw the conclusion that GA optimizes the weights and bias of BP-NN, brings higher converging speed and accurate, and the application of GA-BP algorithm is practicable in neural network prediction model for grain output. Consequently, the neural network model has broad application foreground in the field of ground output prediction by analyzing prediction results.